1 /*
2 * Copyright (C) 2008 The Android Open Source Project
3 *
4 * Licensed under the Apache License, Version 2.0 (the "License");
5 * you may not use this file except in compliance with the License.
6 * You may obtain a copy of the License at
7 *
8 * http://www.apache.org/licenses/LICENSE-2.0
9 *
10 * Unless required by applicable law or agreed to in writing, software
11 * distributed under the License is distributed on an "AS IS" BASIS,
12 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
13 * See the License for the specific language governing permissions and
14 * limitations under the License.
15 */
16 /*
17 * Access .dex (Dalvik Executable Format) files. The code here assumes that
18 * the DEX file has been rewritten (byte-swapped, word-aligned) and that
19 * the contents can be directly accessed as a collection of C arrays. Please
20 * see docs/dalvik/dex-format.html for a detailed description.
21 *
22 * The structure and field names were chosen to match those in the DEX spec.
23 *
24 * It's generally assumed that the DEX file will be stored in shared memory,
25 * obviating the need to copy code and constant pool entries into newly
26 * allocated storage. Maintaining local pointers to items in the shared area
27 * is valid and encouraged.
28 *
29 * All memory-mapped structures are 32-bit aligned unless otherwise noted.
30 */
31 #ifndef _LIBDEX_DEXFILE
32 #define _LIBDEX_DEXFILE
33
34 #include "vm/Common.h" // basic type defs, e.g. u1/u2/u4/u8, and LOG
35 #include "libdex/SysUtil.h"
36
37 /*
38 * gcc-style inline management -- ensures we have a copy of all functions
39 * in the library, so code that links against us will work whether or not
40 * it was built with optimizations enabled.
41 */
42 #ifndef _DEX_GEN_INLINES /* only defined by DexInlines.c */
43 # define DEX_INLINE extern __inline__
44 #else
45 # define DEX_INLINE
46 #endif
47
48 /* DEX file magic number */
49 #define DEX_MAGIC "dex\n"
50 /* version, encoded in 4 bytes of ASCII */
51 #define DEX_MAGIC_VERS "035\0"
52
53 /* same, but for optimized DEX header */
54 #define DEX_OPT_MAGIC "dey\n"
55 #define DEX_OPT_MAGIC_VERS "035\0"
56
57 #define DEX_DEP_MAGIC "deps"
58
59 /*
60 * 160-bit SHA-1 digest.
61 */
62 enum { kSHA1DigestLen = 20,
63 kSHA1DigestOutputLen = kSHA1DigestLen*2 +1 };
64
65 /* general constants */
66 enum {
67 kDexEndianConstant = 0x12345678, /* the endianness indicator */
68 kDexNoIndex = 0xffffffff, /* not a valid index value */
69 };
70
71 /*
72 * access flags and masks; the "standard" ones are all <= 0x4000
73 *
74 * Note: There are related declarations in vm/oo/Object.h in the ClassFlags
75 * enum.
76 */
77 enum {
78 ACC_PUBLIC = 0x00000001, // class, field, method, ic
79 ACC_PRIVATE = 0x00000002, // field, method, ic
80 ACC_PROTECTED = 0x00000004, // field, method, ic
81 ACC_STATIC = 0x00000008, // field, method, ic
82 ACC_FINAL = 0x00000010, // class, field, method, ic
83 ACC_SYNCHRONIZED = 0x00000020, // method (only allowed on natives)
84 ACC_SUPER = 0x00000020, // class (not used in Dalvik)
85 ACC_VOLATILE = 0x00000040, // field
86 ACC_BRIDGE = 0x00000040, // method (1.5)
87 ACC_TRANSIENT = 0x00000080, // field
88 ACC_VARARGS = 0x00000080, // method (1.5)
89 ACC_NATIVE = 0x00000100, // method
90 ACC_INTERFACE = 0x00000200, // class, ic
91 ACC_ABSTRACT = 0x00000400, // class, method, ic
92 ACC_STRICT = 0x00000800, // method
93 ACC_SYNTHETIC = 0x00001000, // field, method, ic
94 ACC_ANNOTATION = 0x00002000, // class, ic (1.5)
95 ACC_ENUM = 0x00004000, // class, field, ic (1.5)
96 ACC_CONSTRUCTOR = 0x00010000, // method (Dalvik only)
97 ACC_DECLARED_SYNCHRONIZED =
98 0x00020000, // method (Dalvik only)
99 ACC_CLASS_MASK =
100 (ACC_PUBLIC | ACC_FINAL | ACC_INTERFACE | ACC_ABSTRACT
101 | ACC_SYNTHETIC | ACC_ANNOTATION | ACC_ENUM),
102 ACC_INNER_CLASS_MASK =
103 (ACC_CLASS_MASK | ACC_PRIVATE | ACC_PROTECTED | ACC_STATIC),
104 ACC_FIELD_MASK =
105 (ACC_PUBLIC | ACC_PRIVATE | ACC_PROTECTED | ACC_STATIC | ACC_FINAL
106 | ACC_VOLATILE | ACC_TRANSIENT | ACC_SYNTHETIC | ACC_ENUM),
107 ACC_METHOD_MASK =
108 (ACC_PUBLIC | ACC_PRIVATE | ACC_PROTECTED | ACC_STATIC | ACC_FINAL
109 | ACC_SYNCHRONIZED | ACC_BRIDGE | ACC_VARARGS | ACC_NATIVE
110 | ACC_ABSTRACT | ACC_STRICT | ACC_SYNTHETIC | ACC_CONSTRUCTOR
111 | ACC_DECLARED_SYNCHRONIZED),
112 };
113
114 /* annotation constants */
115 enum {
116 kDexVisibilityBuild = 0x00, /* annotation visibility */
117 kDexVisibilityRuntime = 0x01,
118 kDexVisibilitySystem = 0x02,
119
120 kDexAnnotationByte = 0x00,
121 kDexAnnotationShort = 0x02,
122 kDexAnnotationChar = 0x03,
123 kDexAnnotationInt = 0x04,
124 kDexAnnotationLong = 0x06,
125 kDexAnnotationFloat = 0x10,
126 kDexAnnotationDouble = 0x11,
127 kDexAnnotationString = 0x17,
128 kDexAnnotationType = 0x18,
129 kDexAnnotationField = 0x19,
130 kDexAnnotationMethod = 0x1a,
131 kDexAnnotationEnum = 0x1b,
132 kDexAnnotationArray = 0x1c,
133 kDexAnnotationAnnotation = 0x1d,
134 kDexAnnotationNull = 0x1e,
135 kDexAnnotationBoolean = 0x1f,
136
137 kDexAnnotationValueTypeMask = 0x1f, /* low 5 bits */
138 kDexAnnotationValueArgShift = 5,
139 };
140
141 /* map item type codes */
142 enum {
143 kDexTypeHeaderItem = 0x0000,
144 kDexTypeStringIdItem = 0x0001,
145 kDexTypeTypeIdItem = 0x0002,
146 kDexTypeProtoIdItem = 0x0003,
147 kDexTypeFieldIdItem = 0x0004,
148 kDexTypeMethodIdItem = 0x0005,
149 kDexTypeClassDefItem = 0x0006,
150 kDexTypeMapList = 0x1000,
151 kDexTypeTypeList = 0x1001,
152 kDexTypeAnnotationSetRefList = 0x1002,
153 kDexTypeAnnotationSetItem = 0x1003,
154 kDexTypeClassDataItem = 0x2000,
155 kDexTypeCodeItem = 0x2001,
156 kDexTypeStringDataItem = 0x2002,
157 kDexTypeDebugInfoItem = 0x2003,
158 kDexTypeAnnotationItem = 0x2004,
159 kDexTypeEncodedArrayItem = 0x2005,
160 kDexTypeAnnotationsDirectoryItem = 0x2006,
161 };
162
163 /* auxillary data section chunk codes */
164 enum {
165 kDexChunkClassLookup = 0x434c4b50, /* CLKP */
166 kDexChunkRegisterMaps = 0x524d4150, /* RMAP */
167
168 kDexChunkReducingIndexMap = 0x5249584d, /* RIXM */
169 kDexChunkExpandingIndexMap = 0x4549584d, /* EIXM */
170
171 kDexChunkEnd = 0x41454e44, /* AEND */
172 };
173
174 /* debug info opcodes and constants */
175 enum {
176 DBG_END_SEQUENCE = 0x00,
177 DBG_ADVANCE_PC = 0x01,
178 DBG_ADVANCE_LINE = 0x02,
179 DBG_START_LOCAL = 0x03,
180 DBG_START_LOCAL_EXTENDED = 0x04,
181 DBG_END_LOCAL = 0x05,
182 DBG_RESTART_LOCAL = 0x06,
183 DBG_SET_PROLOGUE_END = 0x07,
184 DBG_SET_EPILOGUE_BEGIN = 0x08,
185 DBG_SET_FILE = 0x09,
186 DBG_FIRST_SPECIAL = 0x0a,
187 DBG_LINE_BASE = -4,
188 DBG_LINE_RANGE = 15,
189 };
190
191 /*
192 * Direct-mapped "header_item" struct.
193 */
194 typedef struct DexHeader {
195 u1 magic[8]; /* includes version number */
196 u4 checksum; /* adler32 checksum */
197 u1 signature[kSHA1DigestLen]; /* SHA-1 hash */
198 u4 fileSize; /* length of entire file */
199 u4 headerSize; /* offset to start of next section */
200 u4 endianTag;
201 u4 linkSize;
202 u4 linkOff;
203 u4 mapOff;
204 u4 stringIdsSize;
205 u4 stringIdsOff;
206 u4 typeIdsSize;
207 u4 typeIdsOff;
208 u4 protoIdsSize;
209 u4 protoIdsOff;
210 u4 fieldIdsSize;
211 u4 fieldIdsOff;
212 u4 methodIdsSize;
213 u4 methodIdsOff;
214 u4 classDefsSize;
215 u4 classDefsOff;
216 u4 dataSize;
217 u4 dataOff;
218 } DexHeader;
219
220 /*
221 * Direct-mapped "map_item".
222 */
223 typedef struct DexMapItem {
224 u2 type; /* type code (see kDexType* above) */
225 u2 unused;
226 u4 size; /* count of items of the indicated type */
227 u4 offset; /* file offset to the start of data */
228 } DexMapItem;
229
230 /*
231 * Direct-mapped "map_list".
232 */
233 typedef struct DexMapList {
234 u4 size; /* #of entries in list */
235 DexMapItem list[1]; /* entries */
236 } DexMapList;
237
238 /*
239 * Direct-mapped "string_id_item".
240 */
241 typedef struct DexStringId {
242 u4 stringDataOff; /* file offset to string_data_item */
243 } DexStringId;
244
245 /*
246 * Direct-mapped "type_id_item".
247 */
248 typedef struct DexTypeId {
249 u4 descriptorIdx; /* index into stringIds list for type descriptor */
250 } DexTypeId;
251
252 /*
253 * Direct-mapped "field_id_item".
254 */
255 typedef struct DexFieldId {
256 u2 classIdx; /* index into typeIds list for defining class */
257 u2 typeIdx; /* index into typeIds for field type */
258 u4 nameIdx; /* index into stringIds for field name */
259 } DexFieldId;
260
261 /*
262 * Direct-mapped "method_id_item".
263 */
264 typedef struct DexMethodId {
265 u2 classIdx; /* index into typeIds list for defining class */
266 u2 protoIdx; /* index into protoIds for method prototype */
267 u4 nameIdx; /* index into stringIds for method name */
268 } DexMethodId;
269
270 /*
271 * Direct-mapped "proto_id_item".
272 */
273 typedef struct DexProtoId {
274 u4 shortyIdx; /* index into stringIds for shorty descriptor */
275 u4 returnTypeIdx; /* index into typeIds list for return type */
276 u4 parametersOff; /* file offset to type_list for parameter types */
277 } DexProtoId;
278
279 /*
280 * Direct-mapped "class_def_item".
281 */
282 typedef struct DexClassDef {
283 u4 classIdx; /* index into typeIds for this class */
284 u4 accessFlags;
285 u4 superclassIdx; /* index into typeIds for superclass */
286 u4 interfacesOff; /* file offset to DexTypeList */
287 u4 sourceFileIdx; /* index into stringIds for source file name */
288 u4 annotationsOff; /* file offset to annotations_directory_item */
289 u4 classDataOff; /* file offset to class_data_item */
290 u4 staticValuesOff; /* file offset to DexEncodedArray */
291 } DexClassDef;
292
293 /*
294 * Direct-mapped "type_item".
295 */
296 typedef struct DexTypeItem {
297 u2 typeIdx; /* index into typeIds */
298 } DexTypeItem;
299
300 /*
301 * Direct-mapped "type_list".
302 */
303 typedef struct DexTypeList {
304 u4 size; /* #of entries in list */
305 DexTypeItem list[1]; /* entries */
306 } DexTypeList;
307
308 /*
309 * Direct-mapped "code_item".
310 *
311 * The "catches" table is used when throwing an exception,
312 * "debugInfo" is used when displaying an exception stack trace or
313 * debugging. An offset of zero indicates that there are no entries.
314 */
315 typedef struct DexCode {
316 u2 registersSize;
317 u2 insSize;
318 u2 outsSize;
319 u2 triesSize;
320 u4 debugInfoOff; /* file offset to debug info stream */
321 u4 insnsSize; /* size of the insns array, in u2 units */
322 u2 insns[1];
323 /* followed by optional u2 padding */
324 /* followed by try_item[triesSize] */
325 /* followed by uleb128 handlersSize */
326 /* followed by catch_handler_item[handlersSize] */
327 } DexCode;
328
329 /*
330 * Direct-mapped "try_item".
331 */
332 typedef struct DexTry {
333 u4 startAddr; /* start address, in 16-bit code units */
334 u2 insnCount; /* instruction count, in 16-bit code units */
335 u2 handlerOff; /* offset in encoded handler data to handlers */
336 } DexTry;
337
338 /*
339 * Link table. Currently undefined.
340 */
341 typedef struct DexLink {
342 u1 bleargh;
343 } DexLink;
344
345
346 /*
347 * Direct-mapped "annotations_directory_item".
348 */
349 typedef struct DexAnnotationsDirectoryItem {
350 u4 classAnnotationsOff; /* offset to DexAnnotationSetItem */
351 u4 fieldsSize; /* count of DexFieldAnnotationsItem */
352 u4 methodsSize; /* count of DexMethodAnnotationsItem */
353 u4 parametersSize; /* count of DexParameterAnnotationsItem */
354 /* followed by DexFieldAnnotationsItem[fieldsSize] */
355 /* followed by DexMethodAnnotationsItem[methodsSize] */
356 /* followed by DexParameterAnnotationsItem[parametersSize] */
357 } DexAnnotationsDirectoryItem;
358
359 /*
360 * Direct-mapped "field_annotations_item".
361 */
362 typedef struct DexFieldAnnotationsItem {
363 u4 fieldIdx;
364 u4 annotationsOff; /* offset to DexAnnotationSetItem */
365 } DexFieldAnnotationsItem;
366
367 /*
368 * Direct-mapped "method_annotations_item".
369 */
370 typedef struct DexMethodAnnotationsItem {
371 u4 methodIdx;
372 u4 annotationsOff; /* offset to DexAnnotationSetItem */
373 } DexMethodAnnotationsItem;
374
375 /*
376 * Direct-mapped "parameter_annotations_item".
377 */
378 typedef struct DexParameterAnnotationsItem {
379 u4 methodIdx;
380 u4 annotationsOff; /* offset to DexAnotationSetRefList */
381 } DexParameterAnnotationsItem;
382
383 /*
384 * Direct-mapped "annotation_set_ref_item".
385 */
386 typedef struct DexAnnotationSetRefItem {
387 u4 annotationsOff; /* offset to DexAnnotationSetItem */
388 } DexAnnotationSetRefItem;
389
390 /*
391 * Direct-mapped "annotation_set_ref_list".
392 */
393 typedef struct DexAnnotationSetRefList {
394 u4 size;
395 DexAnnotationSetRefItem list[1];
396 } DexAnnotationSetRefList;
397
398 /*
399 * Direct-mapped "anotation_set_item".
400 */
401 typedef struct DexAnnotationSetItem {
402 u4 size;
403 u4 entries[1]; /* offset to DexAnnotationItem */
404 } DexAnnotationSetItem;
405
406 /*
407 * Direct-mapped "annotation_item".
408 *
409 * NOTE: this structure is byte-aligned.
410 */
411 typedef struct DexAnnotationItem {
412 u1 visibility;
413 u1 annotation[1]; /* data in encoded_annotation format */
414 } DexAnnotationItem;
415
416 /*
417 * Direct-mapped "encoded_array".
418 *
419 * NOTE: this structure is byte-aligned.
420 */
421 typedef struct DexEncodedArray {
422 u1 array[1]; /* data in encoded_array format */
423 } DexEncodedArray;
424
425 /*
426 * Lookup table for classes. It provides a mapping from class name to
427 * class definition. Used by dexFindClass().
428 *
429 * We calculate this at DEX optimization time and embed it in the file so we
430 * don't need the same hash table in every VM. This is slightly slower than
431 * a hash table with direct pointers to the items, but because it's shared
432 * there's less of a penalty for using a fairly sparse table.
433 */
434 typedef struct DexClassLookup {
435 int size; // total size, including "size"
436 int numEntries; // size of table[]; always power of 2
437 struct {
438 u4 classDescriptorHash; // class descriptor hash code
439 int classDescriptorOffset; // in bytes, from start of DEX
440 int classDefOffset; // in bytes, from start of DEX
441 } table[1];
442 } DexClassLookup;
443
444 /*
445 * Map constant pool indices from one form to another. Some or all of these
446 * may be NULL.
447 *
448 * The map values are 16-bit unsigned values. If the values we map to
449 * require a larger range, we omit the mapping for that category (which
450 * requires that the lookup code recognize that the data will not be
451 * there for all DEX files in all categories.)
452 */
453 typedef struct DexIndexMap {
454 const u2* classMap; /* map, either expanding or reducing */
455 u4 classFullCount; /* same as typeIdsSize */
456 u4 classReducedCount; /* post-reduction count */
457 const u2* methodMap;
458 u4 methodFullCount;
459 u4 methodReducedCount;
460 const u2* fieldMap;
461 u4 fieldFullCount;
462 u4 fieldReducedCount;
463 const u2* stringMap;
464 u4 stringFullCount;
465 u4 stringReducedCount;
466 } DexIndexMap;
467
468 /*
469 * Header added by DEX optimization pass. Values are always written in
470 * local byte and structure padding. The first field (magic + version)
471 * is guaranteed to be present and directly readable for all expected
472 * compiler configurations; the rest is version-dependent.
473 *
474 * Try to keep this simple and fixed-size.
475 */
476 typedef struct DexOptHeader {
477 u1 magic[8]; /* includes version number */
478
479 u4 dexOffset; /* file offset of DEX header */
480 u4 dexLength;
481 u4 depsOffset; /* offset of optimized DEX dependency table */
482 u4 depsLength;
483 u4 auxOffset; /* file offset of pre-calc auxillary data */
484 u4 auxLength;
485
486 u4 flags; /* some info flags */
487
488 u4 padding; /* induce 64-bit alignment */
489 } DexOptHeader;
490
491 #define DEX_FLAG_VERIFIED (1) /* tried to verify all classes */
492 #define DEX_OPT_FLAG_BIG (1<<1) /* swapped to big-endian */
493 #define DEX_OPT_FLAG_FIELDS (1<<2) /* field access optimized */
494 #define DEX_OPT_FLAG_INVOCATIONS (1<<3) /* method calls optimized */
495
496 #define DEX_INTERFACE_CACHE_SIZE 128 /* must be power of 2 */
497
498 /*
499 * Structure representing a DEX file.
500 *
501 * Code should regard DexFile as opaque, using the API calls provided here
502 * to access specific structures.
503 */
504 typedef struct DexFile {
505 /* directly-mapped "opt" header */
506 const DexOptHeader* pOptHeader;
507
508 /* pointers to directly-mapped structs and arrays in base DEX */
509 const DexHeader* pHeader;
510 const DexStringId* pStringIds;
511 const DexTypeId* pTypeIds;
512 const DexFieldId* pFieldIds;
513 const DexMethodId* pMethodIds;
514 const DexProtoId* pProtoIds;
515 const DexClassDef* pClassDefs;
516 const DexLink* pLinkData;
517
518 /*
519 * These are mapped out of the "auxillary" section, and may not be
520 * included in the file.
521 */
522 const DexClassLookup* pClassLookup;
523 DexIndexMap indexMap;
524 const void* pRegisterMapPool; // RegisterMapClassPool
525
526 /* points to start of DEX file data */
527 const u1* baseAddr;
528
529 /* track memory overhead for auxillary structures */
530 int overhead;
531
532 /* additional app-specific data structures associated with the DEX */
533 //void* auxData;
534 } DexFile;
535
536 /*
537 * Utility function -- rounds up to the nearest power of 2.
538 */
539 u4 dexRoundUpPower2(u4 val);
540
541 /*
542 * Parse an optimized or unoptimized .dex file sitting in memory.
543 *
544 * On success, return a newly-allocated DexFile.
545 */
546 DexFile* dexFileParse(const u1* data, size_t length, int flags);
547
548 /* bit values for "flags" argument to dexFileParse */
549 enum {
550 kDexParseDefault = 0,
551 kDexParseVerifyChecksum = 1,
552 kDexParseContinueOnError = (1 << 1),
553 };
554
555 /*
556 * Correct the byte ordering in a memory-mapped DEX file. This is only
557 * required for code that opens "raw" DEX files, such as the DEX optimizer.
558 *
559 * Return 0 on success.
560 */
561 int dexFixByteOrdering(u1* addr, int len);
562
563 /*
564 * Compute DEX checksum.
565 */
566 u4 dexComputeChecksum(const DexHeader* pHeader);
567
568 /*
569 * Free a DexFile structure, along with any associated structures.
570 */
571 void dexFileFree(DexFile* pDexFile);
572
573 /*
574 * Create class lookup table.
575 */
576 DexClassLookup* dexCreateClassLookup(DexFile* pDexFile);
577
578 /*
579 * Find a class definition by descriptor.
580 */
581 const DexClassDef* dexFindClass(const DexFile* pFile, const char* descriptor);
582
583 /*
584 * Set up the basic raw data pointers of a DexFile. This function isn't
585 * meant for general use.
586 */
587 void dexFileSetupBasicPointers(DexFile* pDexFile, const u1* data);
588
589 /* return the DexMapList of the file, if any */
dexGetMap(const DexFile * pDexFile)590 DEX_INLINE const DexMapList* dexGetMap(const DexFile* pDexFile) {
591 u4 mapOff = pDexFile->pHeader->mapOff;
592
593 if (mapOff == 0) {
594 return NULL;
595 } else {
596 return (const DexMapList*) (pDexFile->baseAddr + mapOff);
597 }
598 }
599
600 /* return the const char* string data referred to by the given string_id */
dexGetStringData(const DexFile * pDexFile,const DexStringId * pStringId)601 DEX_INLINE const char* dexGetStringData(const DexFile* pDexFile,
602 const DexStringId* pStringId) {
603 const u1* ptr = pDexFile->baseAddr + pStringId->stringDataOff;
604
605 // Skip the uleb128 length.
606 while (*(ptr++) > 0x7f) /* empty */ ;
607
608 return (const char*) ptr;
609 }
610 /* return the StringId with the specified index */
dexGetStringId(const DexFile * pDexFile,u4 idx)611 DEX_INLINE const DexStringId* dexGetStringId(const DexFile* pDexFile, u4 idx) {
612 assert(idx < pDexFile->pHeader->stringIdsSize);
613 return &pDexFile->pStringIds[idx];
614 }
615 /* return the UTF-8 encoded string with the specified string_id index */
dexStringById(const DexFile * pDexFile,u4 idx)616 DEX_INLINE const char* dexStringById(const DexFile* pDexFile, u4 idx) {
617 const DexStringId* pStringId = dexGetStringId(pDexFile, idx);
618 return dexGetStringData(pDexFile, pStringId);
619 }
620
621 /* Return the UTF-8 encoded string with the specified string_id index,
622 * also filling in the UTF-16 size (number of 16-bit code points).*/
623 const char* dexStringAndSizeById(const DexFile* pDexFile, u4 idx,
624 u4* utf16Size);
625
626 /* return the TypeId with the specified index */
dexGetTypeId(const DexFile * pDexFile,u4 idx)627 DEX_INLINE const DexTypeId* dexGetTypeId(const DexFile* pDexFile, u4 idx) {
628 assert(idx < pDexFile->pHeader->typeIdsSize);
629 return &pDexFile->pTypeIds[idx];
630 }
631
632 /*
633 * Get the descriptor string associated with a given type index.
634 * The caller should not free() the returned string.
635 */
dexStringByTypeIdx(const DexFile * pDexFile,u4 idx)636 DEX_INLINE const char* dexStringByTypeIdx(const DexFile* pDexFile, u4 idx) {
637 const DexTypeId* typeId = dexGetTypeId(pDexFile, idx);
638 return dexStringById(pDexFile, typeId->descriptorIdx);
639 }
640
641 /* return the MethodId with the specified index */
dexGetMethodId(const DexFile * pDexFile,u4 idx)642 DEX_INLINE const DexMethodId* dexGetMethodId(const DexFile* pDexFile, u4 idx) {
643 assert(idx < pDexFile->pHeader->methodIdsSize);
644 return &pDexFile->pMethodIds[idx];
645 }
646
647 /* return the FieldId with the specified index */
dexGetFieldId(const DexFile * pDexFile,u4 idx)648 DEX_INLINE const DexFieldId* dexGetFieldId(const DexFile* pDexFile, u4 idx) {
649 assert(idx < pDexFile->pHeader->fieldIdsSize);
650 return &pDexFile->pFieldIds[idx];
651 }
652
653 /* return the ProtoId with the specified index */
dexGetProtoId(const DexFile * pDexFile,u4 idx)654 DEX_INLINE const DexProtoId* dexGetProtoId(const DexFile* pDexFile, u4 idx) {
655 assert(idx < pDexFile->pHeader->protoIdsSize);
656 return &pDexFile->pProtoIds[idx];
657 }
658
659 /*
660 * Get the parameter list from a ProtoId. The returns NULL if the ProtoId
661 * does not have a parameter list.
662 */
dexGetProtoParameters(const DexFile * pDexFile,const DexProtoId * pProtoId)663 DEX_INLINE const DexTypeList* dexGetProtoParameters(
664 const DexFile *pDexFile, const DexProtoId* pProtoId) {
665 if (pProtoId->parametersOff == 0) {
666 return NULL;
667 }
668 return (const DexTypeList*)
669 (pDexFile->baseAddr + pProtoId->parametersOff);
670 }
671
672 /* return the ClassDef with the specified index */
dexGetClassDef(const DexFile * pDexFile,u4 idx)673 DEX_INLINE const DexClassDef* dexGetClassDef(const DexFile* pDexFile, u4 idx) {
674 assert(idx < pDexFile->pHeader->classDefsSize);
675 return &pDexFile->pClassDefs[idx];
676 }
677
678 /* given a ClassDef pointer, recover its index */
dexGetIndexForClassDef(const DexFile * pDexFile,const DexClassDef * pClassDef)679 DEX_INLINE u4 dexGetIndexForClassDef(const DexFile* pDexFile,
680 const DexClassDef* pClassDef)
681 {
682 assert(pClassDef >= pDexFile->pClassDefs &&
683 pClassDef < pDexFile->pClassDefs + pDexFile->pHeader->classDefsSize);
684 return pClassDef - pDexFile->pClassDefs;
685 }
686
687 /* get the interface list for a DexClass */
dexGetInterfacesList(const DexFile * pDexFile,const DexClassDef * pClassDef)688 DEX_INLINE const DexTypeList* dexGetInterfacesList(const DexFile* pDexFile,
689 const DexClassDef* pClassDef)
690 {
691 if (pClassDef->interfacesOff == 0)
692 return NULL;
693 return (const DexTypeList*)
694 (pDexFile->baseAddr + pClassDef->interfacesOff);
695 }
696 /* return the Nth entry in a DexTypeList. */
dexGetTypeItem(const DexTypeList * pList,u4 idx)697 DEX_INLINE const DexTypeItem* dexGetTypeItem(const DexTypeList* pList,
698 u4 idx)
699 {
700 assert(idx < pList->size);
701 return &pList->list[idx];
702 }
703 /* return the type_idx for the Nth entry in a TypeList */
dexTypeListGetIdx(const DexTypeList * pList,u4 idx)704 DEX_INLINE u4 dexTypeListGetIdx(const DexTypeList* pList, u4 idx) {
705 const DexTypeItem* pItem = dexGetTypeItem(pList, idx);
706 return pItem->typeIdx;
707 }
708
709 /* get the static values list for a DexClass */
dexGetStaticValuesList(const DexFile * pDexFile,const DexClassDef * pClassDef)710 DEX_INLINE const DexEncodedArray* dexGetStaticValuesList(
711 const DexFile* pDexFile, const DexClassDef* pClassDef)
712 {
713 if (pClassDef->staticValuesOff == 0)
714 return NULL;
715 return (const DexEncodedArray*)
716 (pDexFile->baseAddr + pClassDef->staticValuesOff);
717 }
718
719 /* get the annotations directory item for a DexClass */
dexGetAnnotationsDirectoryItem(const DexFile * pDexFile,const DexClassDef * pClassDef)720 DEX_INLINE const DexAnnotationsDirectoryItem* dexGetAnnotationsDirectoryItem(
721 const DexFile* pDexFile, const DexClassDef* pClassDef)
722 {
723 if (pClassDef->annotationsOff == 0)
724 return NULL;
725 return (const DexAnnotationsDirectoryItem*)
726 (pDexFile->baseAddr + pClassDef->annotationsOff);
727 }
728
729 /* get the source file string */
dexGetSourceFile(const DexFile * pDexFile,const DexClassDef * pClassDef)730 DEX_INLINE const char* dexGetSourceFile(
731 const DexFile* pDexFile, const DexClassDef* pClassDef)
732 {
733 if (pClassDef->sourceFileIdx == 0xffffffff)
734 return NULL;
735 return dexStringById(pDexFile, pClassDef->sourceFileIdx);
736 }
737
738 /* get the size, in bytes, of a DexCode */
739 size_t dexGetDexCodeSize(const DexCode* pCode);
740
741 /* Get the list of "tries" for the given DexCode. */
dexGetTries(const DexCode * pCode)742 DEX_INLINE const DexTry* dexGetTries(const DexCode* pCode) {
743 const u2* insnsEnd = &pCode->insns[pCode->insnsSize];
744
745 // Round to four bytes.
746 if ((((u4) insnsEnd) & 3) != 0) {
747 insnsEnd++;
748 }
749
750 return (const DexTry*) insnsEnd;
751 }
752
753 /* Get the base of the encoded data for the given DexCode. */
dexGetCatchHandlerData(const DexCode * pCode)754 DEX_INLINE const u1* dexGetCatchHandlerData(const DexCode* pCode) {
755 const DexTry* pTries = dexGetTries(pCode);
756 return (const u1*) &pTries[pCode->triesSize];
757 }
758
759 /* get a pointer to the start of the debugging data */
dexGetDebugInfoStream(const DexFile * pDexFile,const DexCode * pCode)760 DEX_INLINE const u1* dexGetDebugInfoStream(const DexFile* pDexFile,
761 const DexCode* pCode)
762 {
763 if (pCode->debugInfoOff == 0) {
764 return NULL;
765 } else {
766 return pDexFile->baseAddr + pCode->debugInfoOff;
767 }
768 }
769
770 /*
771 * Callback for "new position table entry".
772 * Returning non-0 causes the decoder to stop early.
773 */
774 typedef int (*DexDebugNewPositionCb)(void *cnxt, u4 address, u4 lineNum);
775
776 /*
777 * Callback for "new locals table entry". "signature" is an empty string
778 * if no signature is available for an entry.
779 */
780 typedef void (*DexDebugNewLocalCb)(void *cnxt, u2 reg, u4 startAddress,
781 u4 endAddress, const char *name, const char *descriptor,
782 const char *signature);
783
784 /*
785 * Decode debug info for method.
786 *
787 * posCb is called in ascending address order.
788 * localCb is called in order of ascending end address.
789 */
790 void dexDecodeDebugInfo(
791 const DexFile* pDexFile,
792 const DexCode* pDexCode,
793 const char* classDescriptor,
794 u4 protoIdx,
795 u4 accessFlags,
796 DexDebugNewPositionCb posCb, DexDebugNewLocalCb localCb,
797 void* cnxt);
798
799 /* DexClassDef convenience - get class descriptor */
dexGetClassDescriptor(const DexFile * pDexFile,const DexClassDef * pClassDef)800 DEX_INLINE const char* dexGetClassDescriptor(const DexFile* pDexFile,
801 const DexClassDef* pClassDef)
802 {
803 return dexStringByTypeIdx(pDexFile, pClassDef->classIdx);
804 }
805
806 /* DexClassDef convenience - get superclass descriptor */
dexGetSuperClassDescriptor(const DexFile * pDexFile,const DexClassDef * pClassDef)807 DEX_INLINE const char* dexGetSuperClassDescriptor(const DexFile* pDexFile,
808 const DexClassDef* pClassDef)
809 {
810 if (pClassDef->superclassIdx == 0)
811 return NULL;
812 return dexStringByTypeIdx(pDexFile, pClassDef->superclassIdx);
813 }
814
815 /* DexClassDef convenience - get class_data_item pointer */
dexGetClassData(const DexFile * pDexFile,const DexClassDef * pClassDef)816 DEX_INLINE const u1* dexGetClassData(const DexFile* pDexFile,
817 const DexClassDef* pClassDef)
818 {
819 if (pClassDef->classDataOff == 0)
820 return NULL;
821 return (const u1*) (pDexFile->baseAddr + pClassDef->classDataOff);
822 }
823
824 /* Get an annotation set at a particular offset. */
dexGetAnnotationSetItem(const DexFile * pDexFile,u4 offset)825 DEX_INLINE const DexAnnotationSetItem* dexGetAnnotationSetItem(
826 const DexFile* pDexFile, u4 offset)
827 {
828 return (const DexAnnotationSetItem*) (pDexFile->baseAddr + offset);
829 }
830 /* get the class' annotation set */
dexGetClassAnnotationSet(const DexFile * pDexFile,const DexAnnotationsDirectoryItem * pAnnoDir)831 DEX_INLINE const DexAnnotationSetItem* dexGetClassAnnotationSet(
832 const DexFile* pDexFile, const DexAnnotationsDirectoryItem* pAnnoDir)
833 {
834 if (pAnnoDir->classAnnotationsOff == 0)
835 return NULL;
836 return dexGetAnnotationSetItem(pDexFile, pAnnoDir->classAnnotationsOff);
837 }
838
839 /* get the class' field annotation list */
dexGetFieldAnnotations(const DexFile * pDexFile,const DexAnnotationsDirectoryItem * pAnnoDir)840 DEX_INLINE const DexFieldAnnotationsItem* dexGetFieldAnnotations(
841 const DexFile* pDexFile, const DexAnnotationsDirectoryItem* pAnnoDir)
842 {
843 if (pAnnoDir->fieldsSize == 0)
844 return NULL;
845
846 // Skip past the header to the start of the field annotations.
847 return (const DexFieldAnnotationsItem*) &pAnnoDir[1];
848 }
849
850 /* get field annotation list size */
dexGetFieldAnnotationsSize(const DexFile * pDexFile,const DexAnnotationsDirectoryItem * pAnnoDir)851 DEX_INLINE int dexGetFieldAnnotationsSize(const DexFile* pDexFile,
852 const DexAnnotationsDirectoryItem* pAnnoDir)
853 {
854 return pAnnoDir->fieldsSize;
855 }
856
857 /* return a pointer to the field's annotation set */
dexGetFieldAnnotationSetItem(const DexFile * pDexFile,const DexFieldAnnotationsItem * pItem)858 DEX_INLINE const DexAnnotationSetItem* dexGetFieldAnnotationSetItem(
859 const DexFile* pDexFile, const DexFieldAnnotationsItem* pItem)
860 {
861 return dexGetAnnotationSetItem(pDexFile, pItem->annotationsOff);
862 }
863
864 /* get the class' method annotation list */
dexGetMethodAnnotations(const DexFile * pDexFile,const DexAnnotationsDirectoryItem * pAnnoDir)865 DEX_INLINE const DexMethodAnnotationsItem* dexGetMethodAnnotations(
866 const DexFile* pDexFile, const DexAnnotationsDirectoryItem* pAnnoDir)
867 {
868 if (pAnnoDir->methodsSize == 0)
869 return NULL;
870
871 /*
872 * Skip past the header and field annotations to the start of the
873 * method annotations.
874 */
875 const u1* addr = (const u1*) &pAnnoDir[1];
876 addr += pAnnoDir->fieldsSize * sizeof (DexFieldAnnotationsItem);
877 return (const DexMethodAnnotationsItem*) addr;
878 }
879
880 /* get method annotation list size */
dexGetMethodAnnotationsSize(const DexFile * pDexFile,const DexAnnotationsDirectoryItem * pAnnoDir)881 DEX_INLINE int dexGetMethodAnnotationsSize(const DexFile* pDexFile,
882 const DexAnnotationsDirectoryItem* pAnnoDir)
883 {
884 return pAnnoDir->methodsSize;
885 }
886
887 /* return a pointer to the method's annotation set */
dexGetMethodAnnotationSetItem(const DexFile * pDexFile,const DexMethodAnnotationsItem * pItem)888 DEX_INLINE const DexAnnotationSetItem* dexGetMethodAnnotationSetItem(
889 const DexFile* pDexFile, const DexMethodAnnotationsItem* pItem)
890 {
891 return dexGetAnnotationSetItem(pDexFile, pItem->annotationsOff);
892 }
893
894 /* get the class' parameter annotation list */
dexGetParameterAnnotations(const DexFile * pDexFile,const DexAnnotationsDirectoryItem * pAnnoDir)895 DEX_INLINE const DexParameterAnnotationsItem* dexGetParameterAnnotations(
896 const DexFile* pDexFile, const DexAnnotationsDirectoryItem* pAnnoDir)
897 {
898 if (pAnnoDir->parametersSize == 0)
899 return NULL;
900
901 /*
902 * Skip past the header, field annotations, and method annotations
903 * to the start of the parameter annotations.
904 */
905 const u1* addr = (const u1*) &pAnnoDir[1];
906 addr += pAnnoDir->fieldsSize * sizeof (DexFieldAnnotationsItem);
907 addr += pAnnoDir->methodsSize * sizeof (DexMethodAnnotationsItem);
908 return (const DexParameterAnnotationsItem*) addr;
909 }
910
911 /* get method annotation list size */
dexGetParameterAnnotationsSize(const DexFile * pDexFile,const DexAnnotationsDirectoryItem * pAnnoDir)912 DEX_INLINE int dexGetParameterAnnotationsSize(const DexFile* pDexFile,
913 const DexAnnotationsDirectoryItem* pAnnoDir)
914 {
915 return pAnnoDir->parametersSize;
916 }
917
918 /* return the parameter annotation ref list */
dexGetParameterAnnotationSetRefList(const DexFile * pDexFile,const DexParameterAnnotationsItem * pItem)919 DEX_INLINE const DexAnnotationSetRefList* dexGetParameterAnnotationSetRefList(
920 const DexFile* pDexFile, const DexParameterAnnotationsItem* pItem)
921 {
922 return (const DexAnnotationSetRefList*)
923 (pDexFile->baseAddr + pItem->annotationsOff);
924 }
925
926 /* get method annotation list size */
dexGetParameterAnnotationSetRefSize(const DexFile * pDexFile,const DexParameterAnnotationsItem * pItem)927 DEX_INLINE int dexGetParameterAnnotationSetRefSize(const DexFile* pDexFile,
928 const DexParameterAnnotationsItem* pItem)
929 {
930 if (pItem->annotationsOff == 0)
931 return 0;
932 return dexGetParameterAnnotationSetRefList(pDexFile, pItem)->size;
933 }
934
935 /* return the Nth entry from an annotation set ref list */
dexGetParameterAnnotationSetRef(const DexAnnotationSetRefList * pList,u4 idx)936 DEX_INLINE const DexAnnotationSetRefItem* dexGetParameterAnnotationSetRef(
937 const DexAnnotationSetRefList* pList, u4 idx)
938 {
939 assert(idx < pList->size);
940 return &pList->list[idx];
941 }
942
943 /* given a DexAnnotationSetRefItem, return the DexAnnotationSetItem */
dexGetSetRefItemItem(const DexFile * pDexFile,const DexAnnotationSetRefItem * pItem)944 DEX_INLINE const DexAnnotationSetItem* dexGetSetRefItemItem(
945 const DexFile* pDexFile, const DexAnnotationSetRefItem* pItem)
946 {
947 return dexGetAnnotationSetItem(pDexFile, pItem->annotationsOff);
948 }
949
950 /* return the Nth annotation offset from a DexAnnotationSetItem */
dexGetAnnotationOff(const DexAnnotationSetItem * pAnnoSet,u4 idx)951 DEX_INLINE u4 dexGetAnnotationOff(
952 const DexAnnotationSetItem* pAnnoSet, u4 idx)
953 {
954 assert(idx < pAnnoSet->size);
955 return pAnnoSet->entries[idx];
956 }
957
958 /* return the Nth annotation item from a DexAnnotationSetItem */
dexGetAnnotationItem(const DexFile * pDexFile,const DexAnnotationSetItem * pAnnoSet,u4 idx)959 DEX_INLINE const DexAnnotationItem* dexGetAnnotationItem(
960 const DexFile* pDexFile, const DexAnnotationSetItem* pAnnoSet, u4 idx)
961 {
962 return (const DexAnnotationItem*)
963 (pDexFile->baseAddr + dexGetAnnotationOff(pAnnoSet, idx));
964 }
965
966
967 /*
968 * ===========================================================================
969 * Utility Functions
970 * ===========================================================================
971 */
972
973 /*
974 * Retrieve the next UTF-16 character from a UTF-8 string.
975 *
976 * Advances "*pUtf8Ptr" to the start of the next character.
977 *
978 * WARNING: If a string is corrupted by dropping a '\0' in the middle
979 * of a 3-byte sequence, you can end up overrunning the buffer with
980 * reads (and possibly with the writes if the length was computed and
981 * cached before the damage). For performance reasons, this function
982 * assumes that the string being parsed is known to be valid (e.g., by
983 * already being verified). Most strings we process here are coming
984 * out of dex files or other internal translations, so the only real
985 * risk comes from the JNI NewStringUTF call.
986 */
dexGetUtf16FromUtf8(const char ** pUtf8Ptr)987 DEX_INLINE u2 dexGetUtf16FromUtf8(const char** pUtf8Ptr)
988 {
989 unsigned int one, two, three;
990
991 one = *(*pUtf8Ptr)++;
992 if ((one & 0x80) != 0) {
993 /* two- or three-byte encoding */
994 two = *(*pUtf8Ptr)++;
995 if ((one & 0x20) != 0) {
996 /* three-byte encoding */
997 three = *(*pUtf8Ptr)++;
998 return ((one & 0x0f) << 12) |
999 ((two & 0x3f) << 6) |
1000 (three & 0x3f);
1001 } else {
1002 /* two-byte encoding */
1003 return ((one & 0x1f) << 6) |
1004 (two & 0x3f);
1005 }
1006 } else {
1007 /* one-byte encoding */
1008 return one;
1009 }
1010 }
1011
1012 /* Compare two '\0'-terminated modified UTF-8 strings, using Unicode
1013 * code point values for comparison. This treats different encodings
1014 * for the same code point as equivalent, except that only a real '\0'
1015 * byte is considered the string terminator. The return value is as
1016 * for strcmp(). */
1017 int dexUtf8Cmp(const char* s1, const char* s2);
1018
1019
1020 /* for dexIsValidMemberNameUtf8(), a bit vector indicating valid low ascii */
1021 extern u4 DEX_MEMBER_VALID_LOW_ASCII[4];
1022
1023 /* Helper for dexIsValidMemberUtf8(); do not call directly. */
1024 bool dexIsValidMemberNameUtf8_0(const char** pUtf8Ptr);
1025
1026 /* Return whether the pointed-at modified-UTF-8 encoded character is
1027 * valid as part of a member name, updating the pointer to point past
1028 * the consumed character. This will consume two encoded UTF-16 code
1029 * points if the character is encoded as a surrogate pair. Also, if
1030 * this function returns false, then the given pointer may only have
1031 * been partially advanced. */
dexIsValidMemberNameUtf8(const char ** pUtf8Ptr)1032 DEX_INLINE bool dexIsValidMemberNameUtf8(const char** pUtf8Ptr) {
1033 u1 c = (u1) **pUtf8Ptr;
1034 if (c <= 0x7f) {
1035 // It's low-ascii, so check the table.
1036 u4 wordIdx = c >> 5;
1037 u4 bitIdx = c & 0x1f;
1038 (*pUtf8Ptr)++;
1039 return (DEX_MEMBER_VALID_LOW_ASCII[wordIdx] & (1 << bitIdx)) != 0;
1040 }
1041
1042 /*
1043 * It's a multibyte encoded character. Call a non-inline function
1044 * for the heavy lifting.
1045 */
1046 return dexIsValidMemberNameUtf8_0(pUtf8Ptr);
1047 }
1048
1049 /* Return whether the given string is a valid field or method name. */
1050 bool dexIsValidMemberName(const char* s);
1051
1052 /* Return whether the given string is a valid type descriptor. */
1053 bool dexIsValidTypeDescriptor(const char* s);
1054
1055 /* Return whether the given string is a valid reference descriptor. This
1056 * is true if dexIsValidTypeDescriptor() returns true and the descriptor
1057 * is for a class or array and not a primitive type. */
1058 bool dexIsReferenceDescriptor(const char* s);
1059
1060 /* Return whether the given string is a valid class descriptor. This
1061 * is true if dexIsValidTypeDescriptor() returns true and the descriptor
1062 * is for a class and not an array or primitive type. */
1063 bool dexIsClassDescriptor(const char* s);
1064
1065 /* Return whether the given string is a valid field type descriptor. This
1066 * is true if dexIsValidTypeDescriptor() returns true and the descriptor
1067 * is for anything but "void". */
1068 bool dexIsFieldDescriptor(const char* s);
1069
1070 #endif /*_LIBDEX_DEXFILE*/
1071